Background: Evaluation of post procedural hospital length of stay (LoS) in patients considered for TAVI is important for optimization of hospital care. The average LoS for TAVI patients in Germany in 2020 was 11.5 days. The aim of this study was to evaluate predictors of LoS based on data of our department’s TAVI scorecard database.

Method: In this prospective study 299 patients (152 male) aged 81 (IQR: 78-85) years, who underwent a TAVI procedure since 2019 in our institution were enrolled. The primary endpoint of this pilot study was the post TAVI LoS. The TAVI procedure was performed within 48 hours after admission. Patient-reported outcome measures (PROMs, namely Clinical Frailty Scale (CFS), EQ-5D-5L and Kansas City Cardiomyopathy Questionnaire (KCCQ), clinical and biochemical parameters, were assessed at baseline. Post-interventional data, in particular time spent at intensive care unit (TICU) and mobilization time (MT) after TAVI were also analyzed. All variables were analyzed using univariate binary regression. Additionally, all parameters associated with the primary outcome (p<.05) with exception of ES II (score variable) were included in a multivariate analysis for identification of independent predictors. Moreover, a ROC curve analysis of continuous variables was done. The linear predictor score (LPS) describing the probability of a given patient to reach the primary endpoint was the sum of the respective regression coefficients (B) multiplied with the independent predictor values (X). Finally, ROC curve analysis of LPS was performed to obtain its predictive power. The data are presented as numerical values and their median (IQR).

Results: The median LoS was 6 (5-8) days, including 50 (48-72) hours on ICU. MT was 21 (11-26) hours. Therefore, a post TAVI LoS > 6 days was chosen as the primary endpoint. Out of 25 (8.36%) non-survivors 5 died within the first 30 days after TAVI and were excluded from further analysis. Univariate association with endpoint was confirmed for peripheral arterial disease (PAD), CFS, KCCQ, EQ-5D-5L, GFR, NTproBNP, TICU and MT. Multivariate binary regression revealed PAD, CFS, KCCQ, TICU and MT to be independent predictors for post TAVI hospital LoS. Calculated LPS provided an AUC of 0.718. The data are shown on table 1 and figure 1.